Systematic review of the complement components as potential biomarkers of pre-eclampsia: pitfalls and opportunities

The complement system (C) is a crucial component of the innate immune system. An increasing body of research has progressively shed light on the pivotal role of C in immunological tolerance at the feto-maternal interface. Excessive C activation or impaired C regulation may determine the onset of pregnancy-related pathological conditions, including pre-eclampsia (PE). Thus, several studies have investigated the presence of C components or split products in blood matrixes (i.e., plasma, serum), urine, and amniotic fluid in PE. In the current study, we systematically reviewed the currently available scientific literature reporting measurements of C components as circulating biomarkers in PE, based on a literature search using Pubmed, Scopus, and Embase databases. A total of 41 out of 456 studies were selected after full-text analysis. Fourteen studies (34.1%) were identified as measuring the blood concentrations of the classical pathway, 5 (12.1%) for the lectin pathway, 28 (68.3%) for the alternative pathway, 17 (41.5%) for the terminal pathway components, and 16 (39%) for C regulators. Retrieved results consistently reported C4, C3, and factor H reduction, and increased circulating levels of C4d, Bb, factor D, C3a, C5a, and C5b-9 in PE compared to normal pregnancies, depicting an overall scenario of excessive C activation and aberrant C regulation. With evidence of C activation and dysregulation, C-targeted therapy is an intriguing perspective in PE management. Moreover, we also discussed emerging pitfalls in C analysis, mainly due to a lack of experimental uniformity and biased cohort selection among different studies and laboratories, aiming to raise a more comprehensive awareness for future standardization. Systematic review registration https://www.crd.york.ac.uk/prospero/, identifier CRD42024503070.


Introduction
Pre-eclampsia (PE) is a pregnancy-related disorder diagnosed by onset of hypertension (systolic blood pressure ≥ 140 mmHg and/ or diastolic blood pressure ≥ 90 mmHg) after 20 weeks of gestation, associated with proteinuria (≥ 300 mg in 24 hours) and/or presence of kidney or liver dysfunction, neurological complications, hemolysis, or thrombocytopenia, and/or fetal underdevelopment.It is accounted for significant maternal and neonatal morbidity and mortality (1,2).Based on gestational age at clinical presentation, PE is usually classified into preterm (< 37 weeks of gestation), term (≥ 37 weeks of gestation), and post-partum (from 48 hours to 6 days after delivery) (3,4).According to onset, PE has been defined as early-onset PE (EOPE; < 34 weeks) and late-onset PE (LOPE; > 34 weeks) (5).It is now widely accepted that these two entities may differ in etiology and should be accounted for as different phenotypes of the disease (6,7).Conversely, PE classification into mild and severe (SPE) forms has now lost significance.It can also progress to more severe complications, such as HELLP (Hemolysis, Elevated Liver enzymes, and Low Platelet count) syndrome (8) and eclampsia (9).PE affects 5-7% of pregnant women worldwide and can occur in any pregnancy, although several associated risk factors exist.These include PE in previous pregnancies, maternal age (< 20 and ≥ 35 years), metabolic disorders, pregnancies achieved through assisted reproductive technology, and pre-existing conditions, such as chronic kidney disease, thyroid dysfunction, and systemic lupus erythematosus (SLE) (10).A family history of PE is associated with an increased risk, suggesting a genetic background (11).However, no single high-risk gene is associated with PE, reflecting the syndromic nature of the disease.
PE management involves outpatient monitoring of blood pressure and proteinuria, or potential hospitalization in more severe cases.Delivery is often the most effective intervention.Up to now, the soluble fms-like tyrosine kinase 1 (sFlt-1)/placental growth factor (PlGF) ratio is the only predictive biomarker for short-term risk of PE and adverse pregnancy outcomes (12).Additional predictive screenings via algorithms have been extensively evaluated and validated by various institutions (13).Thus, there is a critical need for serum biomarkers for early diagnosis and screening as reliable predictors of PE in the short term.
The etiology of PE is complex and multifactorial.A consistent body of evidence supports the involvement of an abnormal placentation process mainly due to immunological dysfunctions occurring at the fetal-maternal interface (13).Dysregulation of the complement system (C) is recognized as one of the main contributors to PE pathogenesis (14,15).The C is a crucial mediator of the innate immune response, being also involved in cell homeostasis, tissue development and repair, and crosstalk with other endogenous systems (i.e., renin-angiotensin, coagulation, and kinin-kallikrein systems).C consists of more than 50 fluid-phase and membrane-bound proteins.Its activation can be initiated by three distinct pathways (i.e., classical, CP; lectin, LP; and alternative, AP) (Figure 1), converging on the common activation of the major component C3 and in the production of proinflammatory mediators, opsonization, membrane attack complex (MAC) formation, and target cell lysis (16,17).
Balduit et al. 10.3389/fimmu.2024.1419540 Frontiers in Immunology frontiersin.orgrelated pathological conditions, including PE (15).An aberrant or excessive C activation in the placenta can be responsible for PE pathogenesis, leading to inflammation and endothelial dysfunction (14,18,19).This may be mainly due to increased circulating anaphylatoxins, enhanced deposition of C factors in the placenta, or consumption of circulating C factors, leading to placental dysfunction.C activation products can damage endothelial cells, impairing their functions and contributing to the high blood pressure and other symptoms frequently observed in PE (20).This inflammatory condition may also impact the function of other organs, such as liver and kidney, leading to complications.Interestingly, recent research has explored the possibility of targeting the C as a novel therapeutic approach for PE (21,22).C proteins and their activation products have been frequently found in the placenta of women with PE (23), but the relationship between circulating C components and placental C deposition remains elusive.In addition to excessive C activation leading to a simultaneous reduction in C component levels and an increase in activation products, different levels of C components between healthy pregnancies and PE may also be caused by the different hepatic synthesis due to systemic inflammation.Furthermore, one should consider the presence of a massive quantity of circulating microvesicles in the bloodstream of PE patients; they can act as decoys for C components and contribute to an aberrant concentration of C proteins and inhibitors.Thus, research interest has been focused on the evaluation of C components as PE biomarkers before the disease onset (predictive biomarkers) or during the active disease (diagnostic biomarkers).Several studies have investigated the presence of C components or split products in blood matrixes (i.e., plasma, serum), urine, and amniotic fluid in PE.
This systematic review aims to evaluate the current state of the art of C components as potential circulating biomarkers in PE women, provides a general overview of C status in PE, and raises awareness about measuring C components in a reliable and standardized manner.This may help to explore novel predictive approaches and therapeutic opportunities in PE management.

Search methods and data sources
A systematic literature search was carried out to identify original articles reporting the values of circulating C components in PE pregnancies compared to healthy pregnancies.The search strategy was developed following the recommendations of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (24).PubMed, Scopus, and Embase databases were interrogated for relevant articles published until the 13 th March 2024.The search terms included: "preeclampsia", "complement", "serum", "plasma", "blood", and "biomarker".The specific search strategy is reported in the Supplementary Material.The review was registered in the PROSPERO database (CRD42024503070).

Selection of studies
Removal of duplicates and screening of records were performed using Rayyan systematic review tool (25).Two independent investigators (AB and AM) screened the articles by title and abstract, followed by full-text checking for their eligibility.Relevant articles were included if they met all the following eligibility criteria: 1) English language; 2) original peer-reviewed article; and 3) study measuring the levels of at least one C component in the blood of PE vs healthy women.Articles were not included if they were as follows: 1) articles published in languages other than English; 2) case reports/case series; 3) reviews, systematic reviews, meta-analyses, book chapters, conference or comment papers, prospective or retrospective studies; 4) studies reporting biomarkers in urine, cord or fetal blood; 5) studies which did not report numerical values for the measurements of circulating C components; and 6) animal studies.Disparities were resolved by a third investigator (CA).
All considered articles of interest among those identified were reported in this review.

Data extraction
The reported variables include: the first author's name, year of publication, and reference of each study; the values of the biomarkers in PE and control groups, with sample size and units of measurement; p-values and the statistical test used; the time of sample collection and type of blood sample (i.e., serum, plasma); the techniques used for biomarker measurement; notes about EOPE, LOPE, and/or SPE.The recorded statistical measures were reported as mean with standard deviation (mean ± SD) or median with interquartile range [median (IQR)].
Most articles reported a single value for a biomarker, which was included in the study.In a few research articles, multiple values obtained in different cohorts were reported for a single biomarker; in such cases, all cohorts were separately reported.

Studies included for systematic review
The results of the literature search and selection process are illustrated in Figure 2. A total of 812 records were retrieved after the first search in the above-mentioned databases.Duplicates (n = 356) were removed via Rayyan, and the remaining 456 records were screened.During the screening phase, involving the examination of the title and abstract, 330 records were excluded for the following reasons: wrong topic (n = 121), wrong publication type (n = 140), and wrong study design (n = 69).Full-text screening resulted in 126 articles, which were assessed for eligibility.During the eligibility phase, 85 full-text articles were excluded since they did not meet the inclusion criteria: wrong publication type (n = 46), no values were reported (n = 19), wrong topic (n = 7), wrong study design (n = 4), languages other than English (n = 4), and presence of comorbidities (n = 5).Forty-one studies were included in this systematic review.The included papers and relative values are reported in Tables 1-5.
The C markers can be divided into categories based on their involvement in the C cascade.

Lectin pathway components
Our systematic review of currently available research highlighted very few studies (5/41; 12.1%) analyzing blood concentrations of LP components' serum levels, as shown in Table 2. LP components were mainly measured in one study by Larsen and colleagues.They reported significantly higher serum concentrations of CL-L1, Map44, MASP-1, and MASP-3 in PE women than in healthy pregnancies, while significantly lower Hand M-ficolin levels.At the same time, no significant differences were observed in Map19, MASP-2, and MBL (41).MBL levels were further investigated in four additional studies (27,28,30,42), but only one demonstrated an association between increased MBL levels and PE pregnancies (42).

Alternative pathway components
The vast majority of studies focused their interest on the assessment of AP components in serum or plasma samples (28/ 41; 68.3%), including Ba, Bb, C3, C3a, C3c, C3d, Factor B (FB), Factor D (FD), and iC3b, as reported in Table 3. Fragment Ba levels were measured by Blakey and colleagues in two different cohorts of PE patients matched with controls, reporting significantly higher levels of Ba in PE women only in one cohort (31).Seven studies (17.1%) related to Bb concentrations were evaluated, and they

Complement regulators
Several studies (16/41; 39.0%) assessed the circulating levels of different C regulators, mainly focusing on C1-inhibitor and Factor H (FH), as summarized in Table 5.Four studies investigated the association between circulating C1-inhibitor levels and PE (33, 34, 63), but only one reported a significant reduction of C1-inhibitor in women with PE (37).Factor H levels were evaluated in seven studies, which determined an overall reduction of this C regulator in patients with PE (26,28,29,46,47,66) in both EOPE and LOPE (29).A statistically significant reduction in PE cohorts was observed also for Factor I (FI) (67), properdin (31), and vitronectin (68).Velasquez and colleagues reported higher plasma levels of CD59 in PE (64).Conversely, C4b-binding protein (C4BP) (33) and complement receptor 1 (CR1)/CD35 (65) did not show any statistical differences in PE and control groups.

Overview of the complement components in pre-eclampsia
Even though C dysregulation may have a role in affecting placental formation before the disease onset, the contribution of C in PE pathophysiology mainly consists in a secondary mechanism of amplification of tissue injury and inflammation, following endothelial damage and local placental ischemia and hypoxia, and results in a cascade of reactions that contribute to the rapid development and symptom exacerbation of PE at the systemic level (69, 70).When discussing C testing, one should always bear in mind the complexity of C and the interdependence among its individual components.For ease of comprehension, our systematic review dissected the involvement of C components as biomarkers of PE, initially focusing on the contribution of the different C pathways as separate entities, and then moving to a more extensive overview.
As regards the CP and LP, studies were coherent in reporting lower levels of C4 and higher levels of C4d in PE compared to healthy pregnancies.This supports the hypothesis that the relatively low C4 values in PE may indicate a low grade of CP and LP activation with consumption of C4, as suggested by the increase in C4d levels.Agostinis and co-authors also demonstrated that lower levels of C4 were associated with PE (71).Consistently, excessive deposition of C4d at the placental site was frequently reported in PE (23, 31, 72) as a marker of CP and LP activation, even though C4d deposits seem more likely related to CP activation due to the colocalization with C1q (73).The vast majority of the resulting proteolytic cleavage product C4d remains at the activation site, while part of it may eventually enter the circulation as free C4d.
The CP or LP may be triggered by repeated ischemiareperfusion injury and oxidative stress caused by defective placentation in PE (72, 73), while AP exacerbates the activation cycle via an amplification loop.In human PE, C has been shown to be locally activated via the LP triggered by H-ficolin, while MBL did not appear to be particularly involved (23).Larsen and colleagues reported that LP protein concentrations in the serum generally increased throughout the trimesters of normal pregnancy and reduced after delivery in healthy pregnant women and in women with PE (41).A study by Celik and Ozan also reported higher MBL levels in SPE patients compared to women with uncomplicated pregnancies (74), in accordance with further studies (42, 75).Conversely, herein reported results about C1q levels were inconclusive.Dijkstra and colleagues observed no significant differences in serum C1q between PE and control pregnancies in three different cohorts (76),while Agostinis et al. reported a decrease in PE patients, both in serum and plasma samples, maybe due to C1q binding to circulating immunocomplexes or to syncytiotrophoblast extracellular vesicles derived from PE placentas (71, 77, 78).C1q reduction is also supported by the findings of Jia et al., both in EOPE and LOPE (29).In the occurrence of PE related to antibody-mediated autoimmune diseases (e.g., antiphospholipid syndrome), C1q reduction could  associated with CP activation induced by the deposition of pathogenic immunoglobulins in the placenta (79).
The AP is also consistently activated in PE, as demonstrated by the reduction in circulating C3 levels and the increased levels of Ba, Bb, and C3a fragments, being markers of AP activation.Similar results are also consistent with the higher levels of circulating FD in PE (54-57), which is responsible for excessive C activation.Interestingly, the increase in Bb levels is consistent with increased FD levels in PE, considering that FD serves as a serine protease acting on FB to yield the active fragment Bb (57).It is also noteworthy to consider that FD levels may increase in parallel to decreased kidney function and be responsible for enhanced AP activation, particularly in the circulation (80).Elevated circulating levels of C3a in the first trimester of pregnancy were demonstrated as an independent predictive factor for PE (45).Moreover, C3a levels were particularly increased in LOPE (30).Accordingly, local C3 deposition in the placentae of PE patients is significantly higher compared to the control placentae (81).
Evidence about terminal pathway components in PE is rather coherent among studies, reporting increased C5a and C5b-9 levels in the maternal circulation, both in EOPE and LOPE (30, 49).Interestingly, a significant correlation was found between C5b-9 levels and anti-angiogenetic factors (i.e., soluble endoglin and sFlt-1).We should keep in mind that the increase in C5a and C5b-9 may also be directly influenced by additional factors, such as coagulation and fibrinolysis.Thus, urinary excretion of C5b-9 has been proposed as a more valuable marker of SPE than plasma evaluation (51).
Taken together, existing data show that C regulation is also impaired in PE.During normal pregnancy, FH levels usually increase throughout trimesters of healthy pregnancies (28), suggesting that a balanced regulatory response is crucial to avoid excessive systemic AP activation, despite allowing local C3-dependent mechanisms required to support a normal pregnancy.Interestingly, three studies consistently reported that circulating levels of FH were significantly reduced in PE (29,47,66).Yasmin et al. identified lower FH levels in sera collected in the first trimester of pregnancies with PE, proposing FH as a potential predictive biomarker of PE (66).The relative abundance of FH in the third trimester of pregnancy, and particularly the increased FH/C3 ratio, could prevent AP activation, while a reduction in FH may participate in the C dysregulation observed in PE.This scenario is also consistent with an overall reduction of other C regulators, i.e., properdin (31) and FI (67).Reduced concentrations of properdin in maternal plasma allowed to distinguish cases of PE from healthy pregnancies with excellent diagnostic accuracy (31).Excessive AP activity may determine reduced circulating levels of properdin via consumption from tissue deposition.
The overall emerging scenario is an abnormal C activation and regulation in PE.Decreased circulating levels of C1q are frequently observed in combination with lower levels of C4, which may be due to increased consumption of C4 via aberrant activation of the CP or LP.Reduced concentrations of early AP components (i.e., C3, FB, or FH) are also consistent with an extensive depletion of the cascade components due to excessive activation or poor inhibition, while C activation products (e.g., Bb, C3a, C5a, or C5b-9) are increased as a result of the extensive cleavage of C components.A pivotal role in the dysregulation of C activation seems to be carried out by the reduction in FH levels, which could be caused by deficiency or consumption of FH.Interestingly, a recent study by Lokki et al. identified FH variants that predispose to PE onset, confirming the multifactorial nature of the pathophysiological mechanisms underlying PE (82).
Jia and colleagues demonstrated that the best predictive indicator of PE was the combination of five C factors (C1q, Bb, FH, C3, and C4), both in EOPE and LOPE, displaying a potential as diagnostic markers for severe PE (29).

Pitfalls of complement testing in pre-eclampsia
An accurate analysis of a wide range of C components in PE is of utmost importance for prediction and diagnosis, although the reliability of these measurements is often challenging.The main concern of C assessment for diagnosing disease and monitoring therapy is that measured concentrations of C components may widely vary among different laboratories due to a lack of protocol uniformity for pre-analytical sample handling (e.g., collection, processing, and storage), use of different calibrators and techniques (e.g., nephelometry, turbidimetry, and ELISA), or antibodies targeting different epitopes.These pitfalls unveil an urgent need for standardization, in concert with adequate pre-analytical sample handling and storage.However, up to now, performing ELISA with well-defined antibodies has proven to be the most reproducible method for reflecting the actual state of C activation via quantification of Cderived split products next to the native proteins, while nephelometry and turbidimetry do not allow to distinguish between the native nonactivated C proteins and their activated split products (83).
Furthermore, laboratories do not always analyze the proper sample type for C analysis.The use of plasma or serum can lead to different results when certain C factors are considered.In particular, plasma is preferentially chosen for assessing C byproducts to avoid in vitro C activation that occurs during serum preparation due to coagulation and fibrinolysis activation.EDTA-plasma (at least a final concentration of 10 mM), any activation of C is minimized (83).For other factors, the binding of C components to fibrinogen and fibrin in plasma should be taken into account (e.g., C1q) (84).Inconsistences may also be due to sample storage, thawing conditions, and repeated freezing-thawing cycles, which could be crucial variables for specific C components, particularly activation products (85).
Non-uniformity in the timing of sample collection is another issue that needs to be discussed when examining PE, and a serious limitation when comparing the studies collected in this systematic review.Although it may be assumed that the closer the disease onset, the more the potential biomarkers are expected to vary, a different matter is the potential predictive value of the Ccomponents in PE.It is noteworthy that the vast majority of studies gathered in the current review measured C-component expression at the time of PE diagnosis (i.e., in the third trimester of pregnancy).Only a few studies also measured C-components during the first trimester, with limited but encouraging results.He et al. observed significant fluctuations in circulating levels of FB, FH, C1q, C3c, and C4 in PE pregnancies from the first trimester onwards (28).Altered FD and C5a levels in women with PE in the first trimester were also reported (56), as well as FH (66).These preliminary findings emphasize the urgent need to investigate variations in the C components during pregnancy to identify potential predictive markers at early stages.As PE can have multiple causes, it is unlikely that a single timing strategy for biomarker testing can be used to predict all cases of PE (86).Despite requiring a massive effort from clinicians and researchers in terms of number of enrolled women, it is crucial that we embark on prospective longitudinal studies involving blood sample collection at various stages of pregnancy.These studies may allow us to explore the actual time threshold of biomarkers with predictive significance and ensure proper matching for gestational age between healthy and pregnancies with PE.
Diverging results among studies may also be biased by the cohort selection due to different ethnicity, sample size, symptom severity (mild/severe), onset (preterm/term), and presence of comorbidities.For instance, women with African ancestry are at greater risk of PE, likely due to the involvement of specific factors, among which Bb fragment has been accounted (43), while factor B activation has not been observed in specific Caucasian cohorts (33).This evidence suggests that some C factors may serve as predictors of PE only in specific sub-groups of patients, offering a promising avenue for further investigation and potentially explaining the variations among the cohorts analyzed in this review.
Several lines of evidence support the concept that PE should not be considered a single disorder, and the guidelines no longer support the classification into "mild" and "severe" PE.Therefore, the definition of PE severity may not be consistent across studies, based on different updates of the American College of Obstetricians and Gynecologists (ACOG) hypertension in pregnancy guidelines (40,58,59), and may lead to bias in the cohort selection.It is also worth considering that EOPE and LOPE may be explored as separate entities reflecting underlying differences in etiology.The complexity of defining and classifying PE was further highlighted by Than et al., who described four molecular clusters of PE (i.e., canonical/placental, metabolic, immunological, and maternal PE) by proteomic analysis that exhibit distinct clinical phenotypes (87), in line with maternal hemodynamic characteristics of different hypertensive disorders of pregnancy phenotypes (88).
The presence of comorbidities is a crucial factor that must be taken into consideration.Interestingly, several studies that were specifically excluded from the current systematic search report the usefulness of C-component assessment in specific subgroups of patients.For example, predictive value was found for FB and FH serum levels in PE associated with gestational diabetes mellitus (89).

Strengths and limitations of the systematic literature review
To our knowledge, this is the first systematic literature review focusing on currently available research about the assessment of C components as circulating biomarkers of PE.The main strength of the current systematic search includes the comprehensive coverage of available measurements of circulating C components in PE, which were summarized as ready-to-use information.
Nevertheless, we can recognize some methodological limitations.First, not all scientific information may be included in bibliographic databases being reported as unpublished conference papers or congress proceedings.In addition, we included only research papers in which numerical values for the measurement of the C-component were reported, thus possibly neglecting other seminal studies in which numerical values were not explicitly reported.Lastly, only publications in English were included, which may lead to language and country-specific biases, potentially limiting the generalizability of our findings.

Conclusion and perspectives
Since C activation can occur in several scenarios, the specificity of a single C-component test is low, so a panel of tests should be preferred (90).C multiplex ELISA assays, such as panels for simultaneous testing of C activation fragments and intact components or regulators (91), may be particularly advantageous from a technical and reproducibility standpoint.Thus, multiplex assays can help to address the need for a thorough analysis of the status of C activation and regulation and ensure a comprehensive overview of the C cascade.Here, we propose that multiplex C testing should be performed throughout pregnancy to raise awareness of the importance of planning prospective longitudinal studies to identify early predictive markers for PE.
With the mounting evidence of C activation, the potential of Ctargeted therapy for the treatment of PE is a promising avenue.The emergence of new drugs specifically targeting C underscores the growing necessity for soluble biomarkers as companion tests to select and monitor patients who would benefit from these innovative treatments, instilling hope for improved outcomes in the management of PE.

Flow
diagram for article selection according to PRISMA guidelines.